Optimal Design of Single-Cell Experiments within Temporally Fluctuating Environments

Figure 7

Experimental validation of FSP-FIM-based design for optimal biosensor measurements. (a) Distribution of FSP-based MLE estimates for NaCl concentration using the six experimental designs from Figure 6. Each distribution comes from 1,000 replicas of 75 cells per replica spread out over the possible 16 time points. Replica data were sampled randomly from published experimental data [10] that contain two or three biological replicas and 535–4808 cells per time point. The true experimentally applied level of osmotic shock was 0.2 M NaCl. (b) The MLE estimation standard deviation for each experiment design applied to a dataset taken at 0.2 M NaCl (blue). These deviations are compared to FSP-FIM deviation predictions using a piecewise constant model for HOG1 nuclear localization (purple). (c, d) Same as (a, b) but for a true NaCl concentration of 0.4 M.

(a)

(b)

(c)

(d)